CN102346816A - Gene expression profiling for identification of prognostic subclasses in nasopharyngeal carcinomas - Google Patents

Abstract

mRNA transcript profiling can be used to formulate molecular predictors of distant metastasis for primary NPCs. The predicted results are highly correlated with short metastasis-free and overall survival. Predictions are made using 52-genes based and 12- genes based predictors. The invention relates to a method of evaluating and/or predicting the state and the result of the nasopharyngeal carcinoma. The method comprises evaluating the expression profiles of genes related to the nasopharyngeal carcinoma, so the individual prediction and evaluation of the result of a patient having the nasopharyngeal carcinoma can be carried out.

Description

Be used for differentiating the gene expression profile of the prognosis type subclass of nasopharyngeal carcinoma

The application is the application number submitted on 05 22nd, 2008 the dividing an application for the application of " being used for differentiating the gene expression profile of the prognosis type subclass of nasopharyngeal carcinoma " that be 200680043762.3 titles

Present invention relates in general to assess and/or predict the state of nasopharyngeal carcinoma and result's method, comprise the expression of gene level relevant of measuring, make thus and can carry out individuation prediction or assessment these cancer patients' result with this cancer.

Introduction

Nasopharyngeal carcinoma (NPC) is a kind of incidence cancer of unique types, and other malignant tumour of itself and last aerodigestive tract is at epidemiology, pathology, clinical manifestation with to different aspect the reaction of treatment ^1,2The generation that it is believed that NPC is infected relevant with Epstein-Barr virus (EBV) ^{3, 4}Such cancer is local epidemic disease (25-30/100,000/) in southern area of China, Taiwan and south east asia ^5-8It is one of incidence cancer that has more wettability, can invade adjacent organs, invasion and attack pharynxs back and cervical lymph node and propagates in away from the position ^9-11

Radiocurable progress is able to successfully long-term control NPC between three decades in the past ^12-16Undoubtedly, radiotherapy is the main method of treatment NPC.Disease I phase and II phase patient only utilize radiotherapy just can have higher cure rate.Yet the NPC patient of the new diagnosis more than 70% is in the III phase and the IV phase of disease ^14,17These patients need carry out chemotherapy in addition simultaneously to improve its treatment results ^15,16In addition, about 30% disease III phase and IVa/b phase NPC patient produce DISTANT METASTASES IN at last ^14,17, i.e. transfer outside the regional area of NPC.The NPC patient of 10-20% DISTANT METASTASES IN occurs and is in the disease IVc phase when initial diagnosis ^14,17Soon, Most patients is died from this disease after DISTANT METASTASES IN takes place.DISTANT METASTASES IN does not take place under the situation of regional area recurrence appears in III phase and IV phase NPC patient usually, is most important prognosis factor.

Recent meta analysis (meta-analysis) show late period NPC patient to improve life cycle possibly be to carry out the general chemotherapy simultaneously and the result that possibly prevent and control DISTANT METASTASES IN ^15,16Consider that the NPC patient more than 70% is in the III phase and the IV phase of disease, and among these patients of about 30% DISTANT METASTASES IN can take place ^14,17, therefore need when first, successfully differentiate to have the high patient who shifts risk at a distance, thereby further improve these survival time of patients.When this prediction comes true, just can use the methods of treatment of present optimum, can design the methods of treatment that clinical trial testing upgrades then, thereby more effectively to prevent and to control DISTANT METASTASES IN and improve overall life cycle.

The risk of known DISTANT METASTASES IN increases along with the more late stage of NPC ^14,17, used the TNM classification to instruct the appropriate combination of selecting radiation and chemotherapy, to prevent DISTANT METASTASES IN and to improve long term survival ^15,16,31For example, controlling the cancer center at public bright foundation of Taipei guilt and letter has used chemotherapy-radiotherapy (CCRT) to add that the cis-platinum in two cycles and the NACT of 5-FU treat to the III phase patient (having foreclosing of T1N2M0 and T2aN2M0 disease) of AJCC simultaneously ³¹IVa/b patient treats with CCRT, carries out the NACT in 2 cycles subsequently and once in a week with 5-FU and formyl tetrahydrofolic acid maintenance chemotherapy 6 totally months ³¹Although many III phase patients are good and have splendid life cycle (Fig. 6) to this therapeutic response, about 20% III phase patient DISTANT METASTASES IN takes place and life cycle not good.The III phase patient's of DISTANT METASTASES IN overall life cycle and overlapping (Fig. 6) of not having transfer life cycle and IVa/b phase NPC patient take place in 3 years of begin treatment.

Whether these find propose such problem, promptly be in the III phase patient that the DISTANT METASTASES IN risk takes place and should treat more energetically through new adjuvant chemotherapy, renewal or stronger NACT and/or maintenance chemotherapy.Appropriateness to the treatment that is in the IVa/b phase NPC patient that DISTANT METASTASES IN takes place also is a problem (Fig. 6).All these discoveries are stressed when diagnosis and before begin treatment, to identify and are in the clinical importance that the NPC patient in the DISTANT METASTASES IN risk is taken place.It is extremely important for carrying out more effective clinical testing to differentiate that this patient and eliminating are in III phase and IVa/b phase NPC patient's in the low DISTANT METASTASES IN risk ability.

Like previous report, microarray analysis successfully has been used to differentiate the molecular labeling of staging ^18-22And clinical characters ^23-30, for example predict the risk and the overall life cycle of all kinds malignant tumor patient generation DISTANT METASTASES IN.Other research relates to NPC ^37-54

Also referring to the U.S. Patent application No.60/420 of on October 24th, 2002 application; 729; The U.S. Patent application No.60/421 of application on October 25th, 2002; 102; The U.S. Patent application No.60/421 of application on October 25th, 2002; 062; The U.S. Patent application No.60/424 of application on November 8th, 2002; 701; The U.S. Patent application No.60/424 of application on November 8th, 2002; 718; The U.S. Patent application No.60/424 of application on November 8th, 2002; 715; The U.S. Patent application No.60/425 of application on November 12nd, 2002; 256; The U.S. Patent application No.60/448 of application on February 21st, 2003; 462; The U.S. Patent application No.60/448 of application on February 21st, 2003; 466; The U.S. Patent application No.60/457 of application on March 27th, 2003; 877; The U.S. Patent application No.60/458 of application on March 31st, 2003; 373; The U.S. Patent application No.10/291 of application on November 12nd, 2002; 878; The U.S. Patent application No.10/291 of application on November 12nd, 2002; 886; The international patent application No.US02/38216 of application on November 12nd, 2002; And the international patent application No.US02/38222 of application on November 12nd, 2002; The U.S. Patent application No.11/015 of application on Dec 20th, 2004; 764; The U.S. Patent application No.11/090 of application on March 28th, 2005; The U.S. Provisional Application No.60/665 of application on March 25th, 294 and 2005; 652 description, the methodology of using in the said document is incorporated into for referencial use at this.

Summary of the invention

The present invention has differentiated the genome signature that relates to DISTANT METASTASES IN (genomic signatures) among the NPC patient.DISTANT METASTASES IN is taken place NPC patient carrying out forecast method and make the clinician can differentiate the individual patient that is in the excessive risk of improvement can select suitable methods of treatment (for example radiotherapy and/or chemotherapy) to prevent and/or to improve DISTANT METASTASES IN and improve long-term survival rate to them.Therefore, one aspect of the present invention relates to the method that risks and assumptions and clinical effectiveness among the gene expression dose that makes among the NPC patient and the said patient are associated.

Therefore, the present invention relates to a kind of method that the DISTANT METASTASES IN risk takes place the nasopharyngeal carcinoma patient of assessing, said method comprises at least one expression of gene spectrum of listing in the following table 4 and 5 in the said patient's sample of assessment.Preferably, whole basically, that particularly list in whole 52 genes and/or the table 5 whole basically, whole 12 the expression of gene levels particularly in patient's sample, preferred NPC tissue sample, listed in the evaluation form 4.Therefore, can utilize two or more gene in table 4 or 5, the number of the gene of assessing as long as be used for probability analysis is associated with disease result (for example DISTANT METASTASES IN).In other embodiments, for said 12 gene series, can utilize wherein 3 or a plurality of, 4 or a plurality of, 5 or a plurality of, 6 or a plurality of, 7 or a plurality of, 8 or a plurality of, 9 or a plurality of, 10 or a plurality of, perhaps 11 or a plurality of gene.For said 52 gene series; Also can utilize the gene of such number; Perhaps 12 or a plurality of ..., 15 or a plurality of ..., 20 or a plurality of ..., 25 or a plurality of ..., 30 or a plurality of ..., 35 or a plurality of ..., 40 or a plurality of ..., 45 or a plurality of ..., 50 or a plurality of, 51 or a plurality of gene, and at this clear and definite other number gene of statement not.Also can be used to described number gene hereinafter in the combined method of Miao Shuing from each series.Certainly, for Optimization result, utilize usually all or whole basically genes.Therefore, on the other hand, the gene that uses in the preceding method is one or more gene that this paper lists.

The invention still further relates to the set of the whole or inferior collection of these genes relevant in for example medium or kit etc., and the present invention relates to carry out relevant method, medium and the kit that the inventive method is used with DISTANT METASTASES IN.

According to a further aspect in the invention, patient's sample of being analyzed can be any tissue, for example blood, tumour or cell etc.Preferably, said sample is from the NPC tumour.The method that obtains institute's analytic sample is known in the art.

The invention provides and assess or predict the relevant gene sets of DISTANT METASTASES IN among the NPC patient.These genes have the expression pattern relevant with at least a this cancerous phenotype (be horizontal expression or do not have expression).Be appreciated that and also comprise other gene among the NPC.

For the present invention, the biopsy sample that the NPC patient of the clinical data that has complete documentation is certainly collected carries out gene expression pattern research.All biopsy samples of studying are all stored in liquid nitrogen.Only the sample that total RNA is not had an obvious degradation is studied.For the variable relevant with the operator minimized, in the collection of the processing of tumor sample and microarray data, only comprise two technician through senior training.Their random processing similarity number purpose has the patient's of high and low DISTANT METASTASES IN risk sample.The not shown any statistics deviation of statistical analysis (Fig. 7) to the microarray data that undertaken by these two technician and collect.In research process, also use identical fluid automatic transport appearance (fluidic station) and identical scanner.

Minimize with capacity value and the relevant variable of chip processing of chip production, sample preparation, cRNA in order further to make; Use Affymetrix MAS 5.0 softwares that the gene expression intensity data of each microarray is proofreaied and correct and be back-end crop average (trimmed mean) 500, the previous NPC normative reference of in our laboratory, confirming of basis is carried out the fractile correction to the expression intensity of each probe sets subsequently.See that example I is said.The effect of this bearing calibration is confirmed through the GeneChip result who contrasts six NPC sample replications of selecting at random.The result is illustrated in fractile (qauntile) correction program of probe level for proofreading and correct experimental variable really effective (Fig. 8).

Analysis under looking over one's shoulder.As discussed previously, most of NPC patients that DISTANT METASTASES IN does not take place in 3 years of begin treatment have the good long term survival rate, and the NPC patient of existence result badly in 3 years after treatment first DISTANT METASTASES IN takes place usually.These the two groups biopsy samples that are in the patient at disease two ends clinically are used for analyzing, and are used to predict the molecular labeling more reliably of DISTANT METASTASES IN with discriminating.Reported the benefit of taking this method to find reliable molecule predictor (predictor) recently ³⁵

Simultaneously, like Simon etc. ³⁶Said, overmatching (overfitting) is a major defect from a limited number of patient with higher-dimension measured value, finding the classification predictor.In order to overcome the overmatching problem, one group of case of really independently testing that is used to prove conclusively is included, and in this research, comprises patient as much as possible.138 qualified patients are included.In addition, will from 1/3 patient of low-risk and excessive risk group when research has just begun Random assignment to an independent test group (test set) to confirm.Important clinical variable such as age, sex, tumour stage, the duration of following up a case by regular visits to are considered to be used for randomization.All testing group cases all have neither part nor lot in the selection of predicted gene in the training process and confirming of prediction rule.The result of testing group of the present invention shows and is directed against other type entities tumour results reported in sensitivity, specificity and total accuracy and the document quite or better ^23,24,27-30

The tumour cell of survival and the quantity of lymph/inflammation on every side/medium cell can marked changes (biology is heterogeneous) between different patients' biopsy sample.Yet,, therefore the biopsy sample that is used for gene expression spectrum analysis is not carried out histological inspection because the quantity of available biopsy is very limited and the instability of RNA.Equally, the histology of diagnostic biopsy sample and the accuracy onrelevant that predicts the outcome are because be used to diagnose the biopsy site with GeneChip research different.The noise (noise) relevant with the heterogeneity of biopsy sample can limit the accuracy of prediction.

The present invention has produced two preferred prediction rule.See shown in EXAMPLE IV and the V.A characteristic and a k-NN sorting technique based on 52 genes, another is based on 12 genes and logistic regression (logistic regression).

In one group, be used to predict from 52 genes of 9 different SOM bunches.They are summarized in table 4.In these 52 genes, according to NIAID-DAVID Tools ( Http:// apps1.niaid.nih.gov/david/), the member who has participation signal transduction (n=9), mRNA processing (n=7), transcriptional regulatory (n=3), protein synthesis (n=4), nucleotide metabolism (n=4), lipid-metabolism (n=3), protein folding (n=3), cytokinesis (n=2), nuclear translocation (n=2), protein catabolism (n=2), anti-apoptotic (n=1), synthetic (n=1), cell cycle (n=1) of ATP, immune response (n=1), intracellular protein to transport (n=1) and amino acid transport (n=1).The Unknown Function of 7 genes of residue.

In the excessive risk group, there are 22 gene expressions to reduce in these 52 genes, 3 gene expressions increase.When the expression intensity of the predicted gene of contrast excessive risk and low-risk group, notice that the average expression of the gene of nearly all mRNA of participation processing, nuclear translocation, nucleotide metabolism and protein folding in the excessive risk group of prediction is all higher.On the contrary, be in the high group that shifts risk at a distance in prediction, all expression of gene of participating in transcriptional regulatory and protein catabolism all reduce.

For Forecasting Methodology, there are 6 genes also to be present in said 52 genes of predicting through the k-NN method based on 12 genes.According to Affymterix probe sets ID, other 6 genes are not present in the tabulation of said 52 genes.These 12 genes are summarized in table 5.In these 12 genes, there are 3 genes to participate in protein folding, 2 genes are participated in protein synthesis, and 2 genes are participated in biological take place (biogenesis) of ribosomes, and 2 genes are participated in nucleotide metabolism, and 1 gene is participated in mRNA processing and protein catabolism.Last gene (hypothetical protein matter FLJ12671) is known to have kernel exonuclease motif, its actual Unknown Function.The function of all 12 genes all demonstrate with said 52 genes in the function of those genes obviously overlapping.Notice one of 6 genes not being included in said 52 genes (protein that does not contain the POU domain) in fact with said 52 genes in the represented gene of different probe collection ID identical.This gene is participated in mRNA processing.3 genes in the group of said 12 genes are α, β, γ subunits of containing the t-compound 1 (TCP-1) of identical chaperone ³⁴, they participate in tumor cell proliferation.In 12 genes shown in the gene of α of subunit and γ both had been present in, also be present in said 52 genes.The gene of the β of subunit exists only in said 12 genes.All these discoveries illustrate the height consistance between these two groups of predicted genes.

Use the result of prediction rule that the Forecasting Methodology false positive rate low (14% pair 28%) of the Forecasting Methodology of 52 genes than 12 genes is shown.On the contrary, the Forecasting Methodology of 52 genes has higher false negative rate (31% pair 15%) than the Forecasting Methodology of 12 genes.When said Forecasting Methodology is used to select have the high patient who shifts risk at a distance when carrying out clinical testing, false positive and false negative rate need be reduced to minimum value, so that protect low-risk patient's safety and to the correct treatment of high-risk patient.

For example, said two kinds of methods can be combined into a kind of method.Only approve the consistent results of two kinds of methods.Inconsistent result is counted as uncertain result (n=8).See that example VI is said.Through taking this method, false positive rate is reduced to 10%.False negative rate is about 15%.Although uncertain case accounts for all patients' 19% (8/42) in independent test, use the molecule predictor only 2 clinical high-risk patients to be appointed as through combination and belong to uncertain category.By this way, to get rid of the cost of 15% (2/13) clinical high-risk patient, the probability that low-risk patient (good risk patient) is divided in the high-risk patient test group can effectively be minimized mistakenly.Other clinical high-risk patient of 15% is predicted as low-risk patient (table 6) mistakenly.Yet at least 70% high-risk patient will be able to correct discriminated union and can be included in the clinical testing to the high-risk patient design.

Therefore, differentiated the reliable molecules characteristic of 52 genes and 12 genes, be in the excessive risk of generation DISTANT METASTASES IN or the NPC patient of low-risk state with prediction.Said prediction is confirmed through 42 independent test group cases.Be respectively 81%, 76% and 85% by total accuracy of independent test group assessment for said 52 gene expression characteristicses, 12 gene expression characteristicses and assemblage characteristic thereof.These results are suitable with the forecasting research result who also uses the independent test group to confirm who announces recently ^28,30It is reported that prediction breast cancer is 78% to the accuracy rate of new adjuvant chemotherapy reaction ²⁸, the accuracy rate of the lymphatic metastasis of prediction SCCHN is 86% ³⁰Said characterization of molecules can be used for instructing and in NPC patient, carries out new new adjuvant chemotherapy, NACT, maintenance chemotherapy and/or targeted therapy, prevents, improves and control DISTANT METASTASES IN and further improve long-term survival rate being used to.Discriminating with low or high risk NPC patient of DISTANT METASTASES IN also can reduce the excessive or not enough ratio of treatment.

The set of NPC metastasis related gene can be physical set (physical collection) or virtual set (virtual collection).Physical set comprises a group different nucleic acid molecules, and wherein the NPC cancer related gene is present among the said crowd, among the promptly said crowd genome sequence corresponding to the NPC cancer related gene in the said set is arranged, perhaps more commonly, and the nucleic acid molecules of coded sequence.In many embodiments, the sense strand of its corresponding gene of sequence of said nucleic acid molecules is basic identical or identical, and perhaps its corresponding sense strand is complementary, common its corresponding sense strand hybridize under stringent condition.Hybridization conditions (promptly low, medium or high stringent condition) to confirm as those skilled in the art institute known.An example of stringent hybridization condition is hybridization in 0.1SSC (15mM sodium chloride/1.5mM sodium citrate) under 50 ℃ or higher temperature.Another example of stringent hybridization condition is 42 ℃ of incubated overnight in following solution: 50% formamide, 5 * SSC (150mM NaCl, 15mM trisodium citrate); 50mM sodium phosphate (pH7.6); 5 * Denhardt ' s solution, 10% dextran sulfate, and the salmon sperm DNA of 20mg/ml sex change shearing; In 0.1 * SSC, wash filter membrane at about 65 ℃ subsequently.Stringent hybridization condition is the same with above-mentioned condition at least strict hybridization conditions, if wherein the same strict with above-mentioned specific stringent condition about at least 80%, usually about at least 90%, then thinks at least equally strict.Other stringent condition is known in the art, also can be used to differentiate the nucleic acid of this particular embodiment of the present invention.

The nucleic acid of forming said physical set can be strand or double-strandednucleic acid.In addition, the nucleic acid of forming said physical set can be linearity or circular nucleic acid, and each nucleic acid molecules also can comprise other sequence except the NPC cancer related gene, for example the carrier sequence.Various different nucleic acid can be formed said physical set, storehouse for example, and like carrier of the present invention storehouse, wherein the example of dissimilar nucleic acid comprises but the non-DNA of being limited to, cDNA etc. for example, RNA, for example mRNA, cRNA or the like.The nucleic acid of said physical set may reside in the solution or adheres to and promptly is attached on the solid support, on the matrix in the array implement scheme, provides about further describing hereinafter of this different embodiments.

The present invention also provides the virtual set of NPC related gene.Virtual set is meant one or more data file or other mechanized data tissue elements of the sequence information of the gene that comprises set, and wherein said sequence information can be the genome sequence column information, but coded sequence information normally.Said virtual set can be recorded on any computing machine easily or the processor readable storage medium.Computing machine or the processor readable storage medium of having stored said collective data on it can be any media easily, comprise CD, DAT, floppy disk, RAM, ROM etc., and said medium can be read by the hardware components of device.

The present invention also provides NPC Expression of Related Genes spectrum database.This database comprises the express spectra of the various cell/tissues with the relevant phenotype of NPC usually, like each stage of NPC, negative express spectra, prognosis spectrum etc., hereinafter this type spectrum is further described.

Said express spectra and database thereof can be provided in the various media so that its application." medium " is meant the product that contains express spectra information according to the invention.Database of the present invention can be recorded in computer-readable medium and for example can directly be read on any medium with access by computing machine.This medium comprises but the non-magnetic-based storage media that is limited to, like floppy disk, hard disk storage medium and tape; Optical storage medium such as CD-ROM; Electronic storage medium such as RAM and ROM; And the potpourri such as the magnetic/optical storage medium of these storage mediums.Those skilled in the art can be easy to recognize how known any computer-readable medium can be used to produce the product that comprises database information of the present invention at present." record " is meant the process of using any method known in the art canned data on computer-readable medium.Based on the mode that is used for the access stored information, can select any data store organisation easily.Can use various data processor programs and form to store, word processing text (word processing text file) for example, database schema (database format) etc.

As used herein, " computer-based system " is meant hardware, software and the data storage device that is used to analyze information of the present invention.Minimal hardware based on computer system of the present invention comprises a central processing unit (CPU), input media, output unit, and data storage device.The technician can be easy to recognize that any present available computer-based system all is applicable among the present invention.Data storage device can comprise any comprising like above-mentioned recording of information goods of the present invention, memory access device that perhaps can this goods of access.

The many architectures that can use the input and output device are with input and output information in computer-based system of the present invention.A kind of form of output unit is to arrange (rank) to having in various degree the express spectra of similarity with reference expression profile.This technician of being expressed as provides the arrangement of similarity, and the similarity degree that comprises in can the express spectra of differential test.

Gene expression profile can be at a time point determining or at several time point determinings of a period of time.The expression of gene level can be confirmed (for example quantitative polymerase chain reaction (PCR), reverse transcriptase/polymerase PCR) or definite about the method for the quantitative information of gene expression through providing of will develop through any method known in the art.

In another embodiment, gene expression dose is confirmed through quantizing gene expression product such as protein, polypeptide or nucleic acid molecules (for example mRNA, tRNA, rRNA).Quantizing nucleic acid can be through directly quantizing nucleic acid perhaps through corresponding regulatory gene or adjusting sequential element are quantized to carry out.In addition, can the variant of gene be quantized as shearing variant and polymorphism variant.

In another embodiment, gene expression is to measure from the protein of mRNA translation or the level of polypeptide through quantification.The protein or the level of polypeptide and the method that this data are associated with expression are known in the art in the quantification sample.For example, the polyclone or the monoclonal antibody that are specific to protein or polypeptide can obtain through methods known in the art, and are used for detecting and/or the protein or the polypeptide of measuring samples or sample.

In a preferred embodiment, gene expression is to measure through the mRNA level that quantizes in sample or the sample.This can carry out through any method known in the art.In one embodiment, mRNA is contacted with the suitable microarray that comprises the fixing nucleic acid probe that is specific to interested gene, confirm in the sample hybridization degree of probe on the mRNA and microarray.This microarray also within the scope of the invention.The example that produces the method for oligonucleotide microarray is for example being described among the WO 95/11995.This area is easy to know other method.

The gene expression value of mensuration or assessment is the numerical value that derives from the device that can measure gene expression dose.Said numerical value is the raw value that derives from said device, or preferably passes through rescaled, filtration and/or standardized numerical value.See that for example example I is said.Probe hybridization on the nucleic acid (for example mRNA) that has carried out the sample that specific stringent condition handles and the chip.The nucleic acid of analyzing (for example target nucleic acid), the amplification, with hybridization array before with detectable label (for example ³²P or fluorescent marker) carry out mark.After the hybridization, said array insertion can be detected in the scanner of crossing pattern.These patterns detect through the target that detects the mark that adheres to said microarray, if for example said target is fluorescently-labeled, then collect the light from the group of mark, launched as hybridization data.Because the target of mark at suitable stringent condition well known by persons skilled in the art down and the complementary oligonucleotide specific hybrid that comprises in the said microarray; And because the sequence and the position of each oligonucleotides in the known said array, therefore can confirm to be applied to the character of the target nucleic acid of said probe.

The present invention also provides a kind of and has monitored the method for the effect of therapeutic scheme in the individuality through monitoring gene expression profile of the present invention; For example can confirm individual basic gene expression dose, and can during treating, repeat to confirm gene expression profile in different time points.Gene expression profile from the spectrum relevant with not good treatment results change into the relevant spectrum of the treatment results of improving be the effective indication of therapeutic scheme, and repeat represent invalid therapeutic scheme with the relevant express spectra of not good treatment results.

In diagnostic of the present invention is used; Animal (object, the host etc. of said cell/tissue are for example organized and comprise in pair cell or its set; For example mammal such as pet, domestic animal and people etc.) analyze, NPC shifts exists and/or possibility to confirm to take place.So, the diagnostic analysis comprises the method for confirming that this phenotype exists.In certain embodiments, not only confirm the existence of phenotype, but also confirmed the seriousness or the stage of phenotype.In addition, diagnostic methods also comprises the tendency that this cancerous phenotype confirm to take place, and can confirm that thus this cancerous phenotype does not exist but possibly take place.In carrying out said diagnosis and other method,,, carry out method of the present invention then to produce express spectra to deriving from or analyzing derived from the nucleic acid samples of cell, tissue or the object that will be diagnosed.

As stated, analyzing to produce the sample that express spectra is used for said diagnostic method is nucleic acid samples.Said nucleic acid samples comprises nucleic acid a plurality of or that a group is different, and said nucleic acid comprises the NPC Expression of Related Genes information of the interested cell or tissue of diagnosing.Said nucleic acid can comprise RNA or DNA nucleic acid, and for example mRNA, cRNA, cDNA etc. need only the expressing information that said sample keeps its host cell that therefrom obtains or tissue.Can the be known in the art many different modes preparations of said sample, for example through separating mRNA from cell, wherein the mRNA of Fen Liing is increased, and is used to prepare cDNA, cRNA etc., as the differential expression field carry out known.Said sample normally prepares from the cell or tissue that object to be diagnosed is collected; For example collect through using standard scheme to carry out biopsy; The cell or tissue that wherein therefrom can produce this nucleic acid comprises any tissue of the express spectra that wherein has NPC phenotype to be determined, comprises but non-monocyte, endothelium and/or the smooth muscle of being limited to.

Can use any scheme easily from original nucleic acid samples, to produce said express spectra.Though the mode of known various generation express spectra, as be used for those modes in analysis of gene differential expression field, the gene expression profile that a kind of representational scheme that produces gene expression profile easily is based on array produces scheme.Said application is a hybridization analysis, wherein adopts such nucleic acid, said nucleic acid show in the express spectra to be produced each analyzed/" probe " nucleic acid of the gene described.In these are analyzed, at first from the original nucleic acid samples of analyzing, prepare the target nucleic acid sample, wherein preparation can comprise with label (for example signal generation system member) the said target nucleic acid of mark.Behind preparation target nucleic acid sample, sample is contacted under hybridization conditions with said array, thereby with target nucleic acid formation compound with probe sequence that said array surface is adhered to complementation.Then quantitatively or the existence of qualitative detection hybridization complex.The special hybridization technique that can be used for producing the express spectra that adopts in the inventive method is included in incorporates the technology of describing in this paper document for referencial use as follows into: Patent No: 5,143,854,5; 288,644,5,324; 633,5,432,049,5; 470,710,5,492; 806,5,503,980,5; 510,270,5,525; 464,5,547,839,5; 580,732,5,661; 028,5; 800,992, and WO 95/21265, WO96/31622, WO97/10365, WO 97/27317, EP 373 203 and EP 785 280.In these methods, will comprise that its " probe " nucleic acid array of expressing the probe of each NPC related gene of being analyzed contacts with above-mentioned target nucleic acid.Under the for example above-mentioned stringent hybridization condition of hybridization conditions, contact, remove unconjugated nucleic acid then.The hybrid nucleic acid pattern of gained provides each expression of gene information of having detected, and about said gene expression whether and expression levels wherein said expressing information be, and expression data is that express spectra can be quantitatively with qualitatively usually.

In some embodiments, use host, object or patient being diagnosed, lysis and result are predicted about existence, stage or the tendency of the disease that taken place of above-mentioned acquisition about the information of the cell/tissue analyzed.For example, have in the situation of NPC metastatic phenotype at the cell/tissue of confirming to analyze, said information can be used for diagnosing the object that therefrom obtains said cell/tissue to have the possibility of cancer return.

Except the effect of monitoring particular treatment method, the present invention can be used for screening the effect of potential drug candidate in treatment NPC metastatic potential.In this embodiment; Will be with comparing with afterwards sample express spectra before the treatment of said drug candidate, wherein represent the effect of medicine to the transformation of the relevant spectrum of the treatment results of improving through the relevant spectrum of the never good treatment results of gene expression profile of the sample of treatment.Can use conventional method external or in animal model, carry out this analysis.

The Another application of the set of NPC related gene of the present invention is to be used to monitor or evaluate given therapeutic scheme.In this method; Use the patient's of method monitoring experience described herein treatment cell/tissue sample; Wherein express spectra and one or more reference expression profile that obtains compared, whether the disease of treatment is had the influence of hope to confirm given therapeutic scheme.For example, during treating, from the patient, regularly obtain express spectra, itself and a series of references that comprise various NPC transition phase and normal expression spectrum/contrast spectrum are compared.The transformation towards the normal expression spectrum that in the express spectra of monitoring, observes representes that given therapeutic scheme works with the mode of hope.

The therapeutic agent screening is used

The present invention has also been contained discriminating and has been had the method for the material of adjusting (for example strengthening or elimination) NPC metastatic phenotype ability, and said method can be used for differentiating therapeutic agent.

Regulating the discriminating of the compound of this phenotype can use any drug screening technology to realize.The ability of regulating NPC metastatic phenotype decision expression of gene spectrum usually based on said material is analyzed in screening of the present invention.

As used herein, term " material (agent) " is meant any molecule of ability of the biologic activity of the gene outcome that having accords a difference expresses, for example protein, micromolecule or other medicines.The multiple analysis of mixtures that will have different material concentration usually runs parallel, to obtain the reaction to variable concentrations.Usually, one of these concentration are zero-dose or are lower than detection level as negative control.

Candidate substances has contained the material of various chemical classes, but its organic molecule normally, preferred molecular weight greater than 50 less than about 2,500 daltonian little organic compounds.Candidate substances is included as and protein structurally interact necessary functional group, particularly hydrogen bond usually, generally includes at least one amine groups, carbonyl, hydroxyl or carboxyl, preferably at least two said chemical functional groups.Said candidate substances comprises ring-type carbon or heterocycle structure and/or usually with substituted aromatic structure of one or more above-mentioned functions group or fragrant paradigmatic structure.Also in biomolecule, find candidate substances, comprise but non-ly be limited to peptide, sugar, fatty acid, steroids, purine, pyrimidine, derivant, analogue or its combination.

Candidate substances derives from various sources, comprises deriving from synthetic or natural compound.For example, can utilize multiple mode to reach directly synthetic various organic compounds and biomolecule at random, comprise the expression of randomized oligonucleotides and oligopeptides.Perhaps, the native compound library of bacterium, fungi, plant and animal extract (comprising the tissue extract, in order to differentiate the castle's intrinsic factor of the gene outcome that influences differential expression) form can utilize or be easy to and produce.In addition, the natural or library of synthetic generation and chemistry, physics and the biological chemistry mode that compound is easy to through routine are modified, and can be used for producing combinatorial libraries.Can carry out directly or chemical modification at random the known drug preparation, like acidylate, alkanisation, esterification, amidation etc., to produce analogue.

Interested especially candidate substances for example comprises but non-antisense polynucleotides, antibody, the soluble acceptor etc. of being limited to.Antibody and soluble acceptor are interested especially candidate substances, and wherein the gene outcome of target differential expression is cell surface secretion or surperficial (accessible) (for example stablizing related acceptor and other molecule with epicyte) that can reach.

Screening is analyzed can be based on the available and known various technology of those skilled in the art.Normally, said screening analysis comprises that the cell or tissue with the known NPC of having metastatic phenotype contacts with candidate substances, based on evaluating the influence to gene expression profile by the gene expression spectrum analysis of phenotype decision genomic constitution.Can use any method easily to detect said exercising result, in many embodiments, use above-mentioned diagnosis scheme.This analysis is carried out external usually, but can make amendment to be suitable for the body inner analysis to many analyses, for example in the animal cancer model, analyzes.

The drug targeting screening

In another embodiment, the present invention relates to differentiate that gene that this paper lists and product thereof are as the treatment target.Aspect some, this is opposite with the above-mentioned analysis of material of activity that discriminating has adjusting (for example reduce or increase) NPC metastatic phenotype, relates to and differentiates particular phenotype decision gene or its expression product, with as treating target.

In this embodiment, the treatment target is to differentiate through detecting the effect can be proved or be proved the material of regulating NPC phenotype (for example suppress or prevent this phenotype).For example, said material can be the antisense oligonucleotides that is specific to selected genetic transcription thing.For example, said antisense oligonucleotides can have the sequence corresponding to the gene order of listing in this paper table.

The analysis of differentiating the treatment target can utilize the whole bag of tricks well known to those skilled in the art to carry out in every way.For example, contact with known NPC medicament, assess effect and the biologic activity of candidate gene product the NPC phenotype with expressing or crossing the test cell of expressing the gene that candidate gene for example lists in this paper table.The biologic activity of said candidate gene can through test example as to the adjusting (for example increase or the minimizing through transcript level or polypeptide level detects) of the gene expression of coding candidate gene product or to and zymetology or other active adjusting of gene outcome analyze.

The inhibition of NPC metastatic phenotype or prevent and show that said candidate gene product is a kind of suitable treatment target.Can make amendment to be suitable for differentiating the treatment target to analysis that describe among this paper and/or known in the art.Normally, this analysis is carried out external, but can make amendment to be fit to body inner analysis, the for example analysis of in suitable art-recognized animal model, carrying out to many analyses.

Reagent and kit

The present invention also provides reagent and the kit thereof that carries out above-mentioned one or more method.Said reagent and kit thereof can be very different.Interested reagent comprises the reagent that is designed for the above-mentioned NPC phenotype decision of generation expression of gene spectrum especially.

One type this reagent is nucleic acid probe array, wherein has interested NPC metastatic phenotype decision gene.Various array format known in the art has multiple different probe structure, substrate composition and attachment techniques.Interested representative array structure comprises those arrays of incorporating into as follows described in the document for referencial use: U.S. Patent No. 5,143,854,5,288,644,5; 324,633,5,432,049,5,470; 710,5,492,806,5,503,980,5; 510,270,5,525,464,5,547; 839,5,580,732,5,661; 028,5,800,992, and WO 95/21265, WO96/31622, WO 97/10365, WO 97/27317, EP 373 203 and EP 785280.In many embodiments, said array comprises the probe of at least two genes that this paper lists.In certain embodiments, the number gene that exists on the array is at least 5,10,25,50 or more a plurality of, comprises all genes that this paper lists.Said array can only comprise those genes that this paper lists, and perhaps it also can comprise other gene that this paper is unlisted.In certain embodiments, when said array comprised the probe of this other gene, it is about 50% that the number percent of other gene number of existence is no more than, and is no more than about 25% usually.In many embodiments; When comprising this other gene; Most of genes in the said set are NPC cancerous phenotype decision genes; Great majority are meant about at least 75%; Be generally about at least 80%; Sometimes for about at least 85,90,95% or higher, comprise in the said set that 100% gene is the embodiment of NPC cancerous phenotype decision gene.In many embodiments, at least a gene of representing in the said array is its gene that on function, is difficult for the generation of participation NPC cancerous phenotype.

Be particularly suitable for producing the NPC cancerous phenotype and determine that the reagent of the another kind of type of expression of gene spectrum is the set that is designed to the gene-specific primer of this gene of selective amplification.Gene-specific primer and method of application thereof be in U.S. Patent No. 5,994, describes in 076, and said patent is incorporated into for referencial use at this.Interested especially is the gene-specific primer set with primer of at least two kinds of genes that this paper lists.In certain embodiments, the number with this gene of the primer in the said set is at least 5,10,25,50 or more kinds of, comprises all genes that this paper lists.Said gene-specific primer set can only comprise those genes that this paper lists, and perhaps can comprise the primer of other genes that this paper is unlisted.Comprise in the situation of this other gene that in the set of said gene-specific primer it is about 50% that the percentage of said in certain embodiments other gene is no more than, and is no more than about 25% usually.In many embodiments; In the situation that comprises this other gene; Most of genes in the set are NPC phenotype decision genes; Great majority are meant about at least 75%; Be generally about at least 80%; Sometimes for about at least 85,90,95% or higher, comprise in the wherein set that 100% gene is the embodiment of NPC phenotype decision gene.In many embodiments, at least a gene of representative is that it is not easy to participate in the gene that the NPC cancerous phenotype produces on function in the gene-specific primer set.

Kit of the present invention can comprise above-mentioned array and/or gene-specific primer set.Said kit can further comprise one or more the other reagent that is used for the whole bag of tricks; As produce target nucleic acid; The primer of dNTP and/or rNTP; They can be that be pre-mixed or independent; The dNTP of one or more unique tag and/or rNTP; Like dNTP biotinylated or Cy3 or Cy5 mark; Gold or silver particles with different scattering spectrums; Perhaps other synthesize back labelled reagents, like the chemical activity derivant of fluorescent dye, enzyme such as reverse transcriptase; Archaeal dna polymerase; RNA polymerase etc., various buffer mediums, for example hybridization and lavation buffer solution; Prefabricated probe array; The probe purified reagent and the composition of mark, like centrifugal post (spin columns) etc., signal produces and detectable; SA-AP for example, chemiluminescence or chemiluminescent substance etc.

Except mentioned component, said kit further comprises the instructions of putting into practice the inventive method.These instructionss can be present in the said kit by various forms, and wherein one or more can exist in the kit.A kind of form of these instructionss is to be printed on the suitable medium or matrix, for example in the kit package or packing insert one or the plurality of sheets of paper etc. that are printed on said information in the page or leaf etc.Another kind of mode is a computer-readable medium, has for example write down the disk, CD of said information etc. on it.Another kind of mode is the network address that is used for obtaining at remote location through the Internet said information.

The Compounds and methods for that treatment NPC shifts

The present invention also provides the method and composition that can reduce the NPC metastatic potential.The activity that the invention provides through regulating one or more target gene expression or its one or more product alleviates the method for for example treating this disease, and wherein said target gene is the NPC phenotype decision gene that one or more this paper lists.

Some NPC disease part at least too much perhaps exists display abnormality or the highly active gene outcome of mistake by the gene outcome level and causes.So, reduce the level of this gene outcome and/or active will make the recurrence of disease be able to reduce.The technology that reduces expression of target gene level or target gene product activity level has hereinafter been described.

Perhaps, some other NPC disease stage at least partly is owing to lack due to gene expression or gene expression dose reduction or the reduction of gene outcome activity level.So, the activity that increases gene expression dose and/or said gene outcome can be improved said disease.The technology that increases expression of target gene level or target gene product activity level has hereinafter been described.

Suppress the target gene expression of sudden change, synthetic or active compound

As stated, the target gene of participation NPC disease can cause said disease through the increase of target gene activity level.Gene under disease conditions in the cell/tissue is in the up-regulated situation, can utilize various technology to suppress this target gene and/or protein expression, synthetic or active.For example, can use those compounds that present inhibitory activity of differentiating through the method for the invention according to the present invention to improve disease symptoms.As stated, said molecule can comprise but non-little organic molecule, peptide, the antibody etc. of being limited to.The inhibiting antibody technology is described hereinafter.

For example, can give to compete with endogenous ligands the compound of target gene product, wherein said target gene product combines with endogenous ligands.The minimizing of the target gene number that the gained part combines will be regulated endothelial cell physiology.The compound that is used in particular for this purpose comprises for example soluble protein or peptide; As comprise one or more ectodomain of target gene product or the peptide of its part and/or analog; Comprise that (for example U.S. Patent No. 5 is seen in the description that has the fusion of Ig tail about generation to for example soluble fusion like fusion with Ig-tail (Ig-tailed); 116,964.).Perhaps, in conjunction with the target gene product acceptor site but compound such as the ligand analogs or the antibody (for example receptor-ligand antagonist) that do not activate said protein can effectively suppress the target gene product activity.In addition, also can use the antisense and the ribozyme molecule that suppress expression of target gene according to the present invention and suppress unusual target gene activity.This technology is described hereinafter.In addition, also such as hereinafter description, can utilize the unusual target gene of triple helical molecules in inhibiting active.Inhibition antisense thing, ribozyme and triple helical scheme

Show that the compound that reduces the NPC transfer ability is antisense thing, ribozyme and triple helical molecule.This molecule is designed to reduce or suppress the target gene activity of sudden change.The technology that produces and use this molecule is well known to those skilled in the art.

Antisense RNA and dna molecular are through with said target mrna hybridization and stop protein translation and directly block the translation of mRNA.About antisense DNA, preferably derived from translation initiation site, the oligodeoxyribonucleotide between-10 to+10 zones of interested target gene nucleotide sequence for example.Ribozyme is the ribozyme molecule of ability catalysis RNA specificity cracking.The mechanism of action of ribozyme comprises the sequence-specific hybridization of ribozyme molecule and complementary target RNA, with after the endonuclease cutting.The composition of ribozyme molecule must comprise one or more sequence complementary with target gene mRNA, and must comprise the required catalysis sequence of knowing of mRNA cutting.For this reason referring to incorporate 5,093,246 descriptions of U.S. Patent No. for referencial use in full into it.Thus within the scope of the present invention be through engineering approaches hammerhead shape motif ribozymes molecule, it is the endonuclease cutting of the RNA sequence of catalysis coding target gene protein matter specifically and effectively.Specific ribozyme cleavage site in any potential RNA target is to comprise that through scanning the ribozyme cleavage site of the molecule (s) of interest of GUA, GUU and GUC sequence differentiates at first.In case differentiate, then can estimate design feature, as causing the inappropriate secondary structure of oligonucleotide sequence corresponding to the prediction of the short rna sequence of 15-20 the ribonucleotide in the target gene zone that comprises said cleavage site.The adaptability of candidate sequence also can be estimated through using its possibility with the oligonucleotide hybridization of complementation of rnase protection analysis test.The nucleic acid molecules that in triple helical constitutes, is used to suppress to transcribe should be strand and form by deoxyribonucleotide.The basis of these oligonucleotides must be designed to promote triple helical to form through the Hoogsteen basepairing rule, and this needs to exist in the chain of duplex the sequence of sizable one section purine or pyrimidine usually.Nucleotide sequence can will produce the TAT and the CGC+ triplet of three chains that are associated of leap (across) gained triple helical like this based on pyrimidine.The molecule that is rich in pyrimidine provide with a chain of duplex be rich in the base complementrity of purine zone with the direction parallel with this chain.In addition, can select to be rich in the nucleic acid molecules of purine, for example one section sequence that contains the G residue.These molecules form triple helical with the DNA duplex that is rich in the GC base-pair, and wherein most of purine bases are positioned on the chain of target duplex, cause the GGC triplet to cross over three chains in the triple helical.Perhaps, can be used for the potential sequence that triple helical forms by target can increase through generation so-called " switchback " nucleic acid molecules.The Switchback molecule be with replace 5 '-3 ', 3 '-5 ' mode is synthetic; It at first carries out base pairing with a chain of duplex thus; Carry out base pairing with another chain then, thereby need not on a chain of duplex, to exist the sequence of one section sizable purine or pyrimidine.The mRNA that antisense thing described herein, ribozyme and/or triple helical molecule can reduce or suppress normally and the target gene allele of sudden change produces transcribes (triple helical) and/or translates (antisense thing, ribozyme).In order to guarantee to keep the active normal level of target gene; Can be with coding and the nucleic acid molecules of expressing the target gene polypeptide present normal activity through in the gene therapy transfered cell; Said gene therapy those methods as mentioned below do not comprise the sequence to antisense thing, ribozyme or the triple helical treatment susceptible of using.In order to guarantee to keep the target gene activity in basic normal level; Can be with coding and the nucleic acid molecules of expressing the target gene polypeptide present normal activity through in the gene therapy transfered cell as mentioned below, said nucleic acid molecules does not comprise the sequence to the antisense thing of using, ribozyme or triple helical processing susceptible.Perhaps, can preferably normal target gene protein matter be given in the said cell or tissue jointly, to keep the active essential level of cell or tissue target gene.

Can prepare antisense RNA of the present invention and DNA, ribozyme and triple helical molecule according to any method of synthetic DNA known in the art and RNA molecule.These methods comprise chemosynthesis oligodeoxyribonucleotide well known in the art and oligoribonucleotide technology, for example solid phase phosphonic amide (phosphoramidite) chemical synthesis process.Perhaps, transcribe in the external and body of the dna sequence dna that the RNA molecule can be through encoding antisense RNA molecule and produce.This dna sequence dna can mix in the variety carrier that comprises suitable R NA polymerase promoter such as T7 or SP6 polymerase promoter.Perhaps, can be with according to used promoter and the Antisense cDNA construct of composing type or induction type synthesize antisense rna is stablized in the transfered cell system.

Can be to carrying out the various modifications of knowing in the dna molecular to improve born of the same parents' internal stability and half life period.Possible modification comprises but non-being limited in 5 ' and/or 3 ' terminal both sides of said molecule adds ribonucleotide or deoxyribonucleotide sequence, perhaps in the oligodeoxyribonucleotide main chain, uses D2EHDTPA or 2 ' O-methyl key rather than phosphodiester bond.

The antibody of target gene product

Can use and be specific to target gene protein matter and disturb its active antibody to suppress the target gene function.This antibody can use standard technique known in the art to produce to protein self or corresponding to a part of peptide of protein.This antibody comprises but non-polyclonal antibody, monoclonal antibody, Fab fragment, single-chain antibody, the chimeric antibody etc. of being limited to.

In target gene protein matter is in the situation of complete antibody in the born of the same parents, preferably internalization type antibody (internalizing antibody).Yet, can use the lipofectin liposome that the Fab zone fragment of said antibody or combination target gene epi-position is delivered in the cell.In the situation of the fragment of using said antibody, the minimum of the binding structural domain of preferred combination target protein suppresses fragment.For example, can use such peptide, it has the amino acid sequence corresponding to the domain of the antibody variable region that combines target gene protein matter.This peptide can be utilize method well known in the art through chemosynthesis or through recombinant DNA technology generation (for example see Creighton, 1983, as preceding; And Sambrook et al., 1989, as previously mentioned).The strand neutralizing antibody that perhaps, also can combine target gene epi-position in the born of the same parents.This single-chain antibody can for example give through the nucleotide sequence of in the target cell crowd, expressing the coding single-chain antibody; Through for example Marasco et al. (Marasco; W.et al., 1993, Proc.Natl.Acad.Sci.USA 90:7889-7893) said those technology are carried out.

In some cases, target gene protein matter is exoprotein or transmembrane protein.Be specific to one or more ectodomain of gene outcome and disturb its active antibody to be specially adapted to treat breast cancer.It is effective especially because this antibody can directly arrive target region from blood flow.Any technology that gives that being fit to of hereinafter describing gives peptide all can be used for inhibition target gene antibody is given to its site of action effectively.

Recover the active method of target gene

The target gene that causes NPC to shift can be expressed by low under morbid state.Be eliminated by the downward modulation or the activity of target gene product under morbid state at gene, cause can using such method in the situation that disease symptoms produces, it makes the target gene activity level can increase to the level that NPC disease symptoms wherein is improved.The gene activity level can for example perhaps increase through the level that increases the active target gene product that exists through the level that increases the target gene product that exists.

For example, can present the target gene protein matter of the foot part of patient of this symptom with the level of reduction NPC transfer.Can utilize any technology of hereinafter describing to give.Those technology that the known what use is made of of those skilled in the art is hereinafter described are confirmed effective non-toxic dosage concentration of normal target gene protein matter.

In addition, the RNA sequence of coding target gene protein matter directly can be shown the patient that NPC shifts, described RNA sequence gives with the level concentration that is enough to produce the target gene protein matter of improving this symptom.Can utilize hereinafter any technology that gives compound in the realized born of the same parents that describe such as liposome to give method and give this RNA molecule.Said RNA molecule can for example produce through recombinant technique known in the art.

In addition, can be through gene substitution therapy for treating patient.The Gene Partial that the normal target gene protein matter of using carrier one or more copy or the guidance of normal target gene can be had the target gene function produces is inserted in the cell; Said carrier also comprises but non-adenovirus vector, adeno-associated virus vector and the retroviral vector of being limited to except with other particulates such as liposome in the DNA transfered cell.In addition, can utilize like above-mentioned those technology in the normal target-gene sequence importing human body cell.

To contain the cell of the normal target gene of expressing gene sequence then, preferred autogenous cell imports or import in the certain position of patient to improve symptom again.When for example target gene product is born of the same parents' alia gene product of secretion, preferred this cell exchange technology.

Medication preparation and medication

Through differentiate to suppress expression of target gene, synthetic and/or active compound can be treated effective dose and given the patient, with treatment or reduce NPC and shift.The treatment effective dose is meant is enough to make the amount of compound of doing well,improving.

Effective dose

The toxicity of said compound and result of treatment can be confirmed in cell culture or animal used as test through standard drug method, for example confirm LD50 (50% lethal quantity) and ED50 (50% treatment effective dose).Dosage rate between toxicity and the therapeutic action is a therapeutic index, and it can be represented by the LD50/ED50 ratio.The compound that shows higher therapeutic index is preferred.When use presents the compound of toxic side effects, should carefully design induction system with this targeting compounds affected tissue position, make that the potential damage to unaffected cell minimizes, thereby reduce spinoff.

The data that derive from cell culture analysis and zooscopy can be used for formulating the dosage range that is used for human body.The dosage of this compound preferably includes the circulation composition scope of ED50, the very low or avirulence in said concentration range toxicity.Said dosage can change in this scope according to the formulation of using and method of administration.For any compound that uses in the inventive method, the treatment effective dose at first can estimation from cell culture is analyzed.Can in animal model, prepare medicament to reach such circulating plasma concentration range, comprising the scope of the IC50 (promptly reaching the test compounds concentration of the half largest inhibition of symptom) that in cell culture, confirms with par.This information can be used for confirming more accurately the effective dose in human body.Level in the blood plasma can for example be passed through the high performance liquid chroma-tography commercial measurement.

Preparation and application

The pharmaceutical composition of using according to the present invention can use one or more physiology acceptable carrier or excipient to prepare in a usual manner.

Therefore, can prepare acceptable salt of said compound and physiology thereof and solvate, with through suck or insufflation (through port or nose) or oral, contain clothes, enteron aisle outward or the rectally mode give.

For the orally give mode; Pharmaceutical composition can be tablet or the capsule of for example preparing through the acceptable excipient of usual manner medicament; Said excipient such as bond (for example cornstarch, polyvinylpyrrolidone or the hydroxypropyl methylcellulose of preparatory gelatinization), filler (for example lactose, microcrystalline cellulose or calcium monohydrogen phosphate), lubricant (for example dolomol, talcum or silica), disintegrant (for example farina or carboxyrnethyl starch sodium (sodium starch glycolate)), perhaps humidizer (for example lauryl sodium sulfate).Tablet can encapsulate through method well known in the art.Oral flowing product can be for example solution, syrup or form of suspension, perhaps can be dry product, and water or other suitable carriers are set up before using.This flowing product can pass through the suitable additive preparation of usual manner medicament; Said adjuvant such as suspending agent (for example sorbitol syrups, cellulose derivative or hydrogenation edible fat), emulsifying agent (for example lecithin or gum arabic), non-aqueous solution carrier (the for example vegetable oil of apricot kernel oil, grease, ethanol or fractionation), and antiseptic (for example methyl or propyl group-p-metagin or sorbic acid).Said goods also can suitably contain buffer salt, flavouring, pigment and sweetener.

Oral products can suitably be prepared the release with the control reactive compound.For containing the administration of clothes mode, composition can be tablet or the lozenge form of preparing in a usual manner.For the suction administration; The compound that the present invention uses gives with the aerosol form routine from pressurized package or sprayer, uses suitable propellant such as dichlorodifluoromethane, Arcton 11, dichlorotetra-fluoroethane, carbon dioxide or other suitable gas.In the pressurised aerosol situation, dosage unit can be confirmed through the valve control of conveying and metering quantity.Be used for the anther sac (Capsules) of the example gel of inhalator or insufflator and the form that cartridge case (cartridges) can be formulated as the mixed-powder that contains said compound and suitable powder matrix such as lactose or starch.

Compound can be formulated as and is used for giving outside enteron aisle through injection, for example injects or lasting infusion and giving.The injection preparation can exist by unit dosage form, for example is present in the ampoule or multi-dose container of the antiseptic with interpolation.Composition can be suspending liquid, solution or the form of emulsion in oil phase or the aqueous phase carriers, can contain preparaton such as suspending agent, stabilizing agent and/or spreading agent.Perhaps, active component can be the powder type of before using, setting up with suitable carrier such as aseptic apirogen water.

Compound also can be formulated as the per rectum form of medication, like suppository or retention enema, for example contains conventional suppository base such as cupu oil or other glyceride.

Except aforesaid preparation, compound also can be formulated as depot (depot) goods.This long-acting preparation can be through implanting (for example subcutaneous or intramuscular) or giving through intramuscular injection.Therefore, for example compound can perhaps be formulated as the derivant of indissoluble, for example the salt of indissoluble with suitable polymers or hydrophobic material or ion exchange resin preparation.

If desired, then composition may reside in the packing or divider that comprises one or more dosage form that contains active component.Said packing can for example comprise metal or plastic sheeting, as aluminum-plastic packaged.Said packing or divider can attach the administration instructions.

The accompanying drawing summary

Can recognize various characteristic of the present invention and the advantage of following better in conjunction with accompanying drawing, wherein same or analogous part marks with similar fixed reference feature in the different views:

Fig. 1 shows the strategy general introduction that is used for producing and confirming based on mRNA transcript spectrum data the molecule predictor.

Fig. 2 shows the hierarchical cluster analysis (hierarchical cluster analysis) of using 798 genes being selected from 96 training groups (training set) case to carry out.Each gene in a row illustrates in left side and right side.Each case illustrates with post up.The result shows that gene gathers in 6 groups, in the left side there to be the vitta post to illustrate.

Fig. 3 illustrates and analyzes be predicted as the Kaplan-Meier that nothing that low DISTANT METASTASES IN risk and the high case that shifts risk carry out shifts survival probability and overall survival probability at a distance through 52 gene expression characteristicses and k-arest neighbors sorting technique (k-nearest neighbors classifying method).Shown in the result derive from independently testing group case.The p value is calculated with the log-rank check.

Fig. 4 illustrates that nothing that low DISTANT METASTASES IN risk and the high case that shifts risk carry out shifts the probability of existence and the Kaplan-Meier of the probability of overall existence analyzes to being predicted as through 12 gene expression characteristicses and Logic Regression Models at a distance.Shown in the result derive from independently testing group case.The p value is calculated with the log-rank check.

Fig. 5 illustrates the probability and the totally Kaplan-Meier analysis of the probability of existence of combined prediction result to there not being transfer existence according to 52 genes and 12 gene expression characteristicses.The case that inconsistent results is shown between two kinds of characteristics is considered to " uncertain case " (table 6).The case that consistent results is shown is categorized as the classification that low or high distant place shifts risk.When nothing transfer that contrasts low-risk and excessive risk characteristic group and overall life cycle, p value difference＜0.0001 and 0.002.Do not having the life cycle (p=0.09 and 0.05) of transfer and totally do not having significant difference between life cycle (p=0.31 and 0.09) between uncertain group and low-risk or the excessive risk group.The p value is calculated with the log-rank check.

Fig. 6 illustrates having and not having between III phase or the IVa/b phase NPC patient of DISTANT METASTASES IN overall life cycle and analyze with the Kaplan-Meier that does not have the probability that shifts life cycle.The Most patients that comprises in this research began to receive treatment according to the scheme of formulating between 1997 to 2002, regularly carried out subsequently.Above one group of overall life cycle of curve that four groups of patients are shown.Have 106 III phase patients that DISTANT METASTASES IN does not take place, DISTANT METASTASES IN takes place in 27 III phase patients.Similarly, have 47 IV phase patients that DISTANT METASTASES IN does not take place, DISTANT METASTASES IN takes place in 35 IV phase patients.The III phase of DISTANT METASTASES IN or the difference of the overall life cycle between the IVa/b phase patient take place and do not take place all is significant, p value＜0.0001.(p=0.39) taking place or do not taking place does not have significant difference between III phase and the IVa/b phase patient of (p=0.35) DISTANT METASTASES IN.Below one group of III phase that last generation DISTANT METASTASES IN is shown and IVa/b phase patient's nothing shift the probability of life cycle.All III phase patients only treat with chemotherapy-radiotherapy simultaneously and NACT, and IVa/b phase patient is added maintenance chemotherapy.

Fig. 7 illustrates the median intensity correlativity of the probe sets that " has (present) " between two different operating persons.All NPC samples are by two operator's random operations in the training group.24 are not had the case of transfer and 14 positive cases that shift in the operator A operation training group.35 of operator B operations do not have the case of transfer and 18 positive cases that shift.Calculate the median of normalized expression intensity of each probe sets of the case of each operator operation.The result illustrates diagonal angle linear dependence completely, and there be not the systematic bias relevant with the operator in prompting.

Fig. 8 is illustrated in the fractile of probe level correction experimental bias and proofreaies and correct (Quantile normalization) result.CRNA sample from 6 (I-VI) different N PC samples is divided into two parts, with two kinds of U133-A genetic chips in the hybridization of same date not.For each case, at the intensity correlation analysis of all probe sets of the enterprising pedestrian's gene of U133-A genetic chip, as shown in the figure.Above one group of correlativity that illustrates from the probe sets intensity of not standardized chp file.Middle groups is illustrated in the correlativity that the back-end crop average that converts in proportion is a probe sets intensity after 500.Below one group be illustrated in the expression intensity of everyone probe sets proofreaied and correct the correlativity for probe sets intensity after the predetermined standard value through fractile.The result is illustrated in the horizontal fractile correction of probe sets and has effectively proofreaied and correct experimental bias.

Need not describe in further detail, it is believed that those skilled in the art describe according to preamble can at utmost to use the present invention.The present invention that has been illustration only of following preferred particular, and do not have the meaning that limits the present invention by any way.

In preceding addressing following embodiment, unless otherwise indicated, then all temperature all are set at degree centigrade, and all parts and number percent are weight ratio.

In aspect preferred, the invention provides:

1. the method for the risk of DISTANT METASTASES IN takes place in assessment nasopharyngeal carcinoma patient, and said method comprises that assessment derives from least one expression of gene spectrum in the table 4 and 5 listed genes in said patient's the sample.

2. as 1 described method, comprise two or more expression of gene spectrum in 52 genes that evaluation form 4 lists.

3. like 1 described method, it comprises 52 expression of gene spectrums that evaluation form 4 is listed.

4. like 3 described methods, wherein the assessment of his-and-hers watches 4 listed 52 expression of gene is to use the regression model of each gene cluster in 9 gene clusters shown in the table 4 to carry out.

5. like 4 described methods, wherein Lip river base mark (logit score) is to produce to the corresponding regression model formula shown in each the gene cluster use table 1 in said 9 gene clusters.

6. as 5 described methods, wherein the prediction rule of DISTANT METASTASES IN risk is that the k-arest neighbors sorting technique of the said Lip river base mark through being applied to said 9 gene clusters produces.

7. as 1 described method, comprise two or more expression of gene spectrum in 12 genes that evaluation form 5 lists.

8. like 1 described method, comprise 12 expression of gene spectrums that evaluation form 5 is listed.

9. as 8 described methods, wherein the assessment of 12 expression of gene listing of table 5 is to use Logic Regression Models to carry out.

10. like 9 described methods, wherein base mark in Lip river is to use the regression model formula of table 2 to compose and produce based on said 12 expression of gene, and said Lip river base mark is associated with the risk that DISTANT METASTASES IN takes place.

11. like 10 described methods, wherein the prediction rule of low DISTANT METASTASES IN risk is:

12. like 6 described methods, the DISTANT METASTASES IN risk that wherein will derive from said prediction rule compares with the DISTANT METASTASES IN risk that derives from second independent prediction rule, said second prediction rule is to use following formula to assess said risk:

The wherein Lip river base mark regression model formula that is to use table 2 and table 5 12 expression of gene spectrums generation of listing.

13. as 12 described methods, when the DISTANT METASTASES IN risk of from these two kinds of methods, confirming is when low or high, be low-risk or excessive risk with this risk record respectively then; When said definite risk is inconsistent, be uncertain then with this risk record.

14. like 1 described method, wherein said express spectra is assessed in the NPC tumor sample.

15. like 1 described method, the generation from mRNA transcribes of wherein said express spectra.

16. be used for confirming the nucleic acid microarray of nasopharyngeal carcinoma patient DISTANT METASTASES IN risk, it mainly is made up of the probe of confirming following expression of gene spectrum: (a) 52 genes listing of table 4, (b) perhaps (c) said 52 and 12 genes of 12 genes listing of table 5.

17. the set in medium or kit form, all 12 genes that all 52 genes that it is mainly listed by table 4 and/or table 5 are listed; And/or the subclass of the subclass of said 52 genes or said 12 genes or aforementioned two sub-set form, and can effectively predict all that in subclass described in every kind of situation the risk of DISTANT METASTASES IN takes place the nasopharyngeal carcinoma patient.

Embodiment

Embodiment 1

138 biopsy samples to former NPC of acquisition before treatment carry out mRNA transcript analysis of spectrum.Said sample is represented opposite clinical group of two of NPC patient: behind begin treatment, take place in 3 years the DISTANT METASTASES IN group (the excessive risk group, n=47), with no DISTANT METASTASES IN group in the time of examining more than 3 years (the low-risk group, n=91).2/3 of 138 samples are confirmed as the training group at random, and 1/3 confirms as testing group at random.Exercise supervision analysis (Supervised analyses) to disclose in the training group (n=96) and the relevant allelic expression of DISTANT METASTASES IN takes place.The molecule predictor that is disclosed is proved conclusively in the testing group of 42 independent samples.

The patient selects and tumor tissues

Can be utilized in the FF biopsy sample that the public bright foundation of Taipei guilt and letter control 375 NPC patients' that collected in 1992 to 2004 in tumour storehouse, cancer center (KF-SYSCC) primary tumo(u)r and carry out total RNA extraction.Obtain all patients' written Informed Consent Form, and this research obtains the permission of Ethics Committee.105 patients' RNA severely degrade is got rid of it from this research.In remaining 270 samples, only 138 following arbitrary standards of samples met select to be used for this research.To first choice criteria of waiting until the patient who selects for use be its do not take place DISTANT METASTASES IN and 3 years of begin treatment or more for many years in regularly with examining.This group patient (n=91) is called the clinical low-risk group of DISTANT METASTASES IN.Being it to second choice criteria of waiting until the patient who selects for use in accepting to have taken place when treating first DISTANT METASTASES IN or at begin treatment 3 years DISTANT METASTASES IN takes place.DISTANT METASTASES IN is meant the transfer of NPC in lung, bone, liver, kidney, brain and other internal organs.Confirm to shift through histology through fine needle aspiration biopsy or aspiration biopsy.This group patient (n=47) is called the clinical excessive risk group of DISTANT METASTASES IN.

At the primary tumo(u)r biopsy sample of collecting all qualified patients between nineteen ninety-five to 2004 year, get rid of a sample of collecting in 1992.Most of samples (83%) were collected to calendar year 2001 in 1998.Patient age the time is 11 to 71 years old in diagnosis, and intermediate value and mean value are respectively 44 and 45 years old.At initial diagnosis and branch after date, the patient is treated according to draft scheme ³¹The intermediate value of follow-up treatment time and mean value were respectively 3.34 and 3.29.During follow-up treatment, 29 deaths, wherein 28 patients are clinical excessive risk group membership.According to the 1997AJCC regulation NPC patient is carried out by stages.

The research of mRNA transcript analysis of spectrum

(Invitrogen, Carlsbad CA), instruct the total RNA of separation from tissue freezing in liquid nitrogen according to manufacturer to use Trizol reagent.(Qiagen, Valencia CA) are further purified the RNA that separates, and (Agilent Technologies, Waldbronn analyze qualitative through RNA 6000 Nano in Germany) at Agilent 2100 biological analysers to use RNAEasy Mini kit.Be used for all RNA samples of gene expression profile research RNA molecule integrality index (RNA Integrity Number, RIN) all between 6.0 to 10.0 (7.8 ± 1.1, between the mean value ± SD).According to the Affymetrix scheme from total RNA, prepare hybridization target, and with Affymetrix U133A gene chip hybridization.Said U133 A genetic chip contains 22,238 probe sets of about 13,000 human body genes.Said array be characterized as common that kind, for example in the Affymetrix website ( Www.affymetrix.com/products/arrays) describe in detail.In brief, synthetic double chain cDNA from the total RNA of 8 μ g of each sample.Biotin labeled complementary RNA (cRNA) produces through in-vitro transcription from cDNA.The said cRNA of purifying and before hybridization through the chemical mode fragmentation.According to the smart DNA of fragmentation cRNA, probe array contrast, bovine serum albumin(BSA) and the Pacific herring of manufacturer's scheme combination specified quantitative with the preparation potpourri.Said cRNA potpourri and oligonucleotide probe were hybridized 16 hours at 45 ℃ on the U133A genetic chip.After hybridization, use EukGE WS2v4 scheme is washed the probe array of hybridizing automatically in Affymetrix genetic chip washing workstation 400 and is dyeed.In Affymetrix GeneArray scanner 2500, scan the U133A genetic chip afterwards.

The conversion of microarray data and standardization

Use Affymetrix Microarray Analysis Suite (MAS) 5.0 softwares, convert through target correction average 500 and confirm each expression of gene intensity.Carry out number conversion with radix 2 through all human body expression of gene intensity that convert on the U133A genetic chip, use fractile bearing calibration standardization ³²The normative reference that fractile is proofreaied and correct is confirmed from the U133A genetic chip data of the NPC of 164 former NPC, 15 normal nasopharyngeal portion tissues and 23 transfers in our laboratory in advance.The USSN 11/015,764 and the USSN 11/090,294 on March 28th, 2004 that see on Dec 20th, 2004 are said, and said document is incorporated into for referencial use at this with its full content.The observational measurement data of other U133A genetic chip data comprise number percent as " existences " and perceptible gene (52.1 ± 5.8%, mean value ± SD), reach GAPDH 3 ' and 5 ' ratio (0.96 ± 0.18, mean value ± SD).These two parameters are all supported the good oeverall quality of sample and analysis.

Example II

The statistical analysis of data

The statistical analysis method of using is summarized in Fig. 1 and is illustrated.

Sample is divided into training group and testing group

In 138 NPC cases that our research institute comprises, any DISTANT METASTASES IN does not take place and with examining above 3 years in 91 cases after beginning treatment first.These 91 cases are categorized as clinical low DISTANT METASTASES IN risk group.In remaining 47 cases, perhaps in 3 years after beginning treatment first DISTANT METASTASES IN takes place all in treatment first.It is categorized as the clinical high group that shifts risk at a distance.For all patients, diagnosis date and to treat equispaced between the date first be 20 ± 51 days (mean value ± SD) first.Use SAS software (version 9.1) with the patient of each risk group 2/3 be appointed as at random the training group (SAS software can derive from SAS Institute Inc., Cary, NC.).1/3 patient is appointed as testing group.(I and II phase are to III and IV phase), successive treatment time (≤and＞4.5 years) are told level by stages according to sex, age (≤with＞45 years old), 1997AJCC TNM to low-risk group patient.Excessive risk group patient is reached from (≤with＞1.5 years) classification of treatment first to the time that DISTANT METASTASES IN takes place according to sex, age, TNM by stages.62 low-risk cases and 34 excessive risk cases are arranged in the training group; Independently in the testing group 29 low-risk cases and 13 excessive risk cases are being arranged.Said independently testing group sample is not participated in training process.

The gene that selection is used to analyze

Only select to analyze through the probe of Affymetrix MAS5.0 software definite " existence " its expression in all NPC samples of getting rid of the testing group sample.At first use GeneLinker Platinum 4.5 softwares (Predictive Patterns Software; Inc.; Inverary Canada), carries out the Kruskal-Wallis check analysis through standardization with to the expression data of number conversion to each gene between low-risk group and excessive risk group.Select 798 genes of p value＜0.05 further to study through self-organization mapping (SOM) method (GeneLinker Platinum 4.5software).

SOM analyzes

Analyze said 798 genes through SOM.The parameter that SOM analyzes comprises Pearson correlation coefficient, and the one dimension reflection (map dimension) of 2 (highly) * 3 (width) of gene direction, distance.For reference vector and algorithm character, Use Defaults.The selection of 2 * 3 dimensions is according to the hierarchical cluster analysis of 798 genes of training group case is carried out (Fig. 2).This method provides selects SOM to analyze the objective way of mapping scope.All 798 genes are divided into six different SOM gene clusters (I to VI).

(Chicago Illinois) further selects the gene in each SOM gene cluster through binary progressive method logistic regression for SPSS, Inc. to use SPSS 9.0 softwares.The inlet of progressive method logistic regression (Entry) and remove (removal) p value and be respectively＜0.05 and＞0.1.After the logistic regression analysis, the number gene that is selected from SOM gene cluster II, V and VI is respectively 6,6 and 5.During logistic regression, SOM gene cluster I, III and IV are separated fully.Subsequently, the gene to SOM gene cluster I, III and IV carries out two-dimentional SOM analysis separately.The gained number of the gene of selecting among SOM gene cluster Ia, Ib, IIIa, IIIb, Iva and the IVb is respectively 9,5,8,4,6 and 3.Therefore, 52 genes are arranged in nine SOM gene clusters.

The establishment of Forecasting Methodology

Establish and differentiate two kinds of Forecasting Methodologies that the high NPC patient who shifts risk is taken place at a distance.First method is based on 52 genes in nine SOM gene clusters.Gene to each gene cluster is set up regression model.The formula of each regression model is listed in table 1.From each formula, produce Lip river base mark (logit score) to each sample.Nine of each sample Lip river base marks (logit score) are used to produce prediction rule in the training group, use SAS 9.0 softwares to carry out through k-arest neighbors (k-NN) sorting technique.Use " k " value of training group independence test 1,3,5,10 and 30.Carry out the Leave-one-out cross validation.According to the leave-one-out cross validation, the k value of check is optimum to be provided at 10 o'clock, selects it to carry out Forecasting Methodology.

Second kind of Forecasting Methodology is based on 12 genes that derive from 197 genes among the original SOM gene cluster I.As stated, advancing method before use select to analyze and to carry out a logistic regression period three SOM gene cluster separation fully is shown.The potential high predicted of these three SOM gene clusters of results suggest is worth.For the gene of differentiating which SOM gene cluster can be used for prediction reliably; Use training group case from each SOM gene cluster, to select gene through binary progressive method logistic regression (SPSS 9.0 softwares), the generation that is chosen in of gene separates fully and correctly stops before.The gene of from each gene cluster, selecting is used to produce Lip river base mark (logit score), uses it for estimated probability.Probability is higher than 0.5 and is appointed as low DISTANT METASTASES IN risk, and probability is lower than 0.5 and is appointed as and highly shifts risk at a distance.The result illustrates 12 genes that use training group case from SOM gene cluster I, to select and produces optimum.Formula based on the regression model of 12 genes among the SOM gene cluster I is shown in table 2.

Analyze life cycle

Using SAS 9.0 softwares not have the life cycle of transfer through Kaplan-Meier log rank check analyzed with overall life cycle.Be interpreted as life cycle treat first Start Date to for the last time with examine between date or date of death during.Do not have to shift be meant life cycle treat first between Start Date to the date of diagnosing DISTANT METASTASES IN first during.

EXAMPLE III

Result's general introduction

As shown, confirm two predictor differentiate be in take place high shift risk at a distance or low DISTANT METASTASES IN risk in NPC patient.First predictor is based on 52 genes and the k-arest neighbors sorting technique in nine different self-organizations mapping cluster numbers (self organizing maps clusters).Second predictor is based on 12 genes and Logic Regression Models.These two kinds of methods all can be predicted testing group case independently short-and-medium DISTANT METASTASES IN interval and short overall life cycle strongly.Total accurate rate of these two kinds of methods assessment in testing group independently is respectively 81% and 76%.When these two kinds of Forecasting Methodologies of combination, accurate rate increases to 85%.The generation DISTANT METASTASES IN risk ratio of the estimation between excessive risk characteristic group and the low-risk characteristic group be 11.1 (95% fiducial interval 2.4-52.4, p=0.002).

Patient's characteristic

In order to differentiate that prediction is in the gene expression characteristics that high or low DISTANT METASTASES IN risk takes place and have the NPC patient of bad or good overall life cycle, analysis exercises supervision between two groups of clinical NPC patients that fully confirm.Be fully recognized that treating the NPC patient that DISTANT METASTASES IN does not take place in back 3 years first and had good long term and do not have and shift life cycle and overall life cycle ^31,33On the contrary, the NPC patient that DISTANT METASTASES IN takes place in when treating first or afterwards 3 years dies from this disease usually and has bad life cycle.In 138 patients of our research, 91 patients belong to clinical low-risk group, and 47 patients belong to clinical excessive risk group.Specifying 2/3 low-risk patient and high-risk patient at random is the training group, and 1/3 patient is a testing group.These patients' characteristic is shown in table 3.The Clinical symptoms of between low-risk group and excessive risk group training group and testing group case, listing does not have significant difference.In addition, being distributed as of the DISTANT METASTASES IN position of training group (n=34) and testing group (n=13): bone 50%vs 69%, liver 50%vs 54%, lung 38%vs 23%, other positions 15%vs 23%.Other positions comprise brain, kidney, spleen and pelvic organ.It is similar to distribute between two groups.

EXAMPLE IV

Signatures to predict DISTANT METASTASES IN through 52 genes

In order to differentiate predicted gene, carry out the height supervision and analyze.Only use after getting rid of the testing group case in all NPC samples the gene (4,814 probe sets) that can confirm as " existence " through Affymetrix MAS 5.0 softwares in this research.Between the clinical low-risk group of training group and clinical excessive risk group, carry out the Kruskal-Wallis check then.Selection illustrates its 798 genes of expressing significant difference (p＜0.05) and further studies.Use 798 gene pairs training group cases selecting to carry out unsupervised hierarchical cluster analysis.The result illustrates six oligogenes bunch (Fig. 2).Based on the hierarchical cluster analysis result, the one dimension of selection 2 * 3 reflection carries out SOM and analyzes.Most of genes in each SOM bunch flock together, and are parallel with the gene cluster that produces through hierarchical cluster analysis.

Gene in each said six SOM bunch gene cluster is further analyzed through logistic regression, in the training group, selects low-risk and excessive risk case with progressive method.Three SOM cluster genes illustrate the problem of separating fully.Every bunch gene is directed against bunch further analysis of two Asias through SOM in these three bunches.From nine SOM gene clusters, select totally 52 genes, be used to produce forecast model to DISTANT METASTASES IN.These 52 genes summaries are shown in Table 4.Be used for from each SOM bunch of gene, confirming that the formula of Lip river base mark (logit score) is shown in table 1.

Therefore, from said 52 expression of gene intensity of each case, produce 9 Lip river base marks (logit score).Said 9 Lip rivers base marks (logit score) of each case are used for setting up prediction rule through the k-NN sorting technique in the training group.To the Leave-one-out cross validation of prediction rule prediction training group case DISTANT METASTASES IN being shown is that high risk specificity and sensitivity are respectively 98% (61/62) and 88% (30/34).Total accurate rate is 95% (91/96).When analyzing independently testing group case (n=42) according to the prediction rule of confirming, specificity, sensitivity and total accurate rate are respectively 86% (25/29), 69% (9/13) and 81% (34/42).(p=0.0007) confirm the robustness (robustness) of said 52 predictive genes factors to the definite probabilistic method of Fisher ' s (exact test) of correlativity (association).With regard to DISTANT METASTASES IN and short overall life cycle; The danger of estimating between the excessive risk characteristic group of prediction and the low-risk characteristic group is respectively 7.1 (p=0.002 than (estimated hazard ratio); 95% fiducial interval 2.2-23.5) and 5.4 (p=0.001,95% fiducial interval 1.5-19.4).

The nothing that is predicted as the patient in low DISTANT METASTASES IN risk (n=29) or high (n=13) testing group that shifts risk at a distance when contrast shifts life cycle during with overall life cycle, is predicted as to have the high patient who shifts risk at a distance and have significantly short nothing and shift life cycle (p=0.0001) and short overall life cycle (p=0.003) (Fig. 3).

EXAMPLE V

Through 12 gene expression characteristics prediction DISTANT METASTASES IN

Second kind of Forecasting Methodology is based on 12 genes through selecting in the gene among SOM bunch of I in the logistic regression analysis.The formula that is used for calculating the probability of low DISTANT METASTASES IN risk is summarized at table 2.Said 12 genes are listed in the table 5.When said 12-predictive genes rule application during in said independently testing group case (n=42), the sensitivity that the prediction excessive risk shifts, specificity and total accuracy are respectively 85% (11/13), 72% (21/29) and 76% (32/42).For these 12 predictive genes factors, to the definite probabilistic method of Fisher ' s (p=0.0008) confirmation of correlativity and the similar robustness of said 52 predictive genes factors.With regard to DISTANT METASTASES IN and short overall life cycle; The danger ratio of estimating between the excessive risk of prediction and the low-risk characteristic group of prediction is respectively 8.2 (p=0.006; 95% fiducial interval 1.8-37.4) and 6.3 (p=0.02,95% fiducial interval 1.3-29.7).

Shift life cycle during with overall life cycle when analyzing the nothing that is predicted as low-risk (n=23) or excessive risk (n=19) case through 12 predictive genes factors, the result illustrates through being predicted as the high patient who shifts risk at a distance to be had significantly short nothing and shifts life cycle (p=0.0007) and overall life cycle (p=0.007) (Fig. 4).

Example VI

Through assemblage characteristic prediction DISTANT METASTASES IN

When consistance in independent test group case of the result of two kinds of Forecasting Methodologies analyzing EXAMPLE IV and V, there are 22 examples (52%) all to be predicted as the low-risk case and 12 examples (29%) are predicted as excessive risk case (table 6) by these two kinds of methods.There are 8 examples (19%) inconsistent between these two kinds of methods, think uncertain case.In the situation of unanimity (n=34), it is error prediction cases that 5 examples are arranged, and 3 examples are that false positive and 2 examples are false negative cases.Therefore, total accurate rate is 85% (29/34).With regard to DISTANT METASTASES IN and short overall life cycle, the danger ratio of the estimation between excessive risk and the low-risk characteristic group is respectively 11.1 (p=0.002,95% fiducial interval 2.4-52.4) and 8.5 (p=0.009,95% fiducial interval 1.7-42.8).To carrying out analyzing life cycle and show that the excessive risk case has obviously relatively poor nothing and shifts life cycle and overall life cycle (Fig. 5) through being predicted as low-risk those cases through being predicted as low-risk, excessive risk and uncertain case.There is not the difference not statistically significant that shifts between life cycle and overall life cycle between low-risk and the uncertain case.

Example VII A

Contrast life cycle between III phase and the IV phase NPC patient

In order to recognize the potential clinical effectiveness of definite Forecasting Methodology, collect 133 III phases and 82 Iva and IVb (IVa/b) phase NPC patient's clinical data.In these patients, only 90 examples are parts of above-mentioned mRNA transcript analysis of spectrum research.All patients are all according to the scheme of drafting ³¹Between 1997 to 2003, treat and follow-up with examining.With TNM III or IVa/b phase patient according to the generation of DISTANT METASTASES IN subsequently and further be divided into two groups.Shift life cycle with nothing the overall life cycle that contrasts these four groups of patients.The result illustrates and shifts life cycle similar (Fig. 6) with the IVa/b phase patient's (n=35) that DISTANT METASTASES IN also takes place overall life cycle with nothing life cycle with the nothing transfer III phase NPC patient's that DISTANT METASTASES IN takes place subsequently overall life cycle.On the contrary, III phase (n=106) or IVa/b phase (n=47) patient that DISTANT METASTASES IN do not take place those patients that DISTANT METASTASES IN takes place have overall life cycle (Fig. 6) preferably.

The result is illustrated in and has two among III phase or the IVa/b phase patient not on the same group.One group presents low-risk level that DISTANT METASTASES IN takes place and good clinical effectiveness.DISTANT METASTASES IN takes place and has bad existence result (Fig. 6) in another group in 3 years of begin treatment.Such viewpoint has been supported in these discoveries; Promptly not only have important prevision meaning, select suitable patient to test new form of therapy to improve the mode of long-term treatment effect but also provide for being in the accurately predicting that high patient in shifting risk is taken place at a distance.

List of references

1.Lo，W.K.，To，K.F.&?Huang，D.P.Focus?on?nasopharyngeal?carcinoma.Cancer?Cell?5，423-428(2004).

2.Altun，M.et?al.Undifferentiated?nasopharyngeal?cancer(UCNT)：current?diagnostic?and?therapeutic?aspects.Int.J.Radiat.Oncol.Biol.Phys.32：859-877(1995).

3.Young，L.S.&?Riskinson，A.B.Epsein-Barr?virus：40years?on.Nature?Rev?Cancer?4：757-768(2004).

4.Cohen，J.I.Epstein-Barr?virus?infection.New?Eng?J?Med?343，481-492(2000).

5.Yu，M.C.Nasopharyngeal?carcinoma：Epidemiology?and?Dietary?Factors?pp.39-47.Lyon：IARC，(1991)

6.Ho，J.Nasopharyngel?carcinoma.Adv.Cancer?Res.15：59-72(1972)

7.Yeh，S.&?Cowdry，E.Incidence?of?malignant?tumors?in?Chinese.Cancer7：425-436，1954.

8.Yu，M.C.&?Henderson，B.E.Nasopharyngeal?cancer.In：Schottenfield?D，Fraumeni?JF，eds.Cancer?epidemiology?and?prevention.2 ^nd?ed.New?York：Oxford?Univ.Press?1996：603-18.

9.Fandi，A.et?al.Nasopharyngeal?cancer：epidemiology，staging，and?treatment.Semin.Oncol.21：382-397(1994).

10.Vikram，B，et?al.Patterns?of?failure?in?carcinoma?of?the?nasopharynx：failure?at?distant?sites.Head?Neck?Surg.8：276-279(1986).

11.Hung，S.C.Nasopharyngeal?cancer：a?review?of?1605?paatients?treated?radically?with?cobalt?60.Int.J.Radiat.Oncol.Biol.Phys.6：401-407(1980).

12.Geara，F.B.etal.Carcinoma?of?the?nasopharynx?treated?by?radiothera[y?alone：determinants?of?distant?metastasis?and?survival.Radiother.Oncol.43：53-61(1997).

13.Perez，C.A.et?al.Carcinoma?of?the?nasopharynx；factors?affecting?prognosis.Int.J.Radiat.Oncol.Biol.Phys.23：271-280(1992).

14.Lee，A.W.M.et?al.Retrospective?analysis?of?5037?patients?with?nasopharyngeal?carcinoma?treated?during?1976-1985：overall?survival?and?patterns?of?failure.Int.J.Radiat.Oncol.Biol.Phys.23：261-270(1992).

15.Agulnik，M.&?Siu，L.L.State-of-the-art?management?of?nasopharyngeal?carcinoma：current?and?future?directions.Br.J.Cancer?92：799-806(2005).

16.Langendijk?JA，Leemans?CR，Buter?J，et?a；/The?additional?value?of?chemotherapy?to?radiotherapy?in?locally?advanced?nasopharyngeal?carcinoma：a?meta-analysis?of?he?published?literature.J?Clin?Oncol?2004；22：4604-4612.

17.Chang，J.T-C.et?al.Nasopharyngeal?carcinoma?staging?by?9180F-fluorodeoxyglucose?positron?emission?tomography.Int.J.Radiat.Oncol.Biol.Phys.62：501-507(2005).

18.Perou，C.M.et?al.Molecular?portraits?of?human?breast?tumours.Nature?406：747-752(2000).

19.Alizadeh?AA，Eisen?MB，Davis?RE，et?al.Distinct?types?of?diffuse?large?B-cell?lymphoma?identified?by?gene?expression?profiling.Nature?2001；405：503-511.

20.Khan?J，Wei?JS，Ringer?M，et?al.Classification?and?diagnostic?prediction?of?cancers?using?gene?expression?profiling?and?artificial?neural?networks.Nature?Med?2001；7；673-679.

21.Dyrskjot?L，Thykjaer?T，Kruhoffer?M，et?al.Identifying?distinct?classes?of?bladder?carcinoma?using?microarrays.Nature?Gent?2003；33：90-96.

22.Bullinger?L，Dohner?K，Blair?e，et?al.Use?of?gene-expression?profiling?to?identify?prognostic?subclasses?in?adult?acute?myeloid?leukemia.New?Eng?J?Med?2004；350：1605-1615.

23.van′t?Veer?LJ，Dai?H，van?de?Vijver?MJ，et?al：Gene?expression?profiling?predicts?clinical?outcome?of?breast?cancer.Nature?2002；415：530-536.

24.Pomeroy?SL，Tamayo?P，Gaasenbeek?M，et?al.Prediction?of?central?nervous?system?embryonal?tumour?outcome?based?on?gene?expression.Nature2002；415：436-442.

25.Shipp?MA，Ross?KN，Tamayo?P，et?al.Diffuse?large?B-cell?lymphoma?outcome?prediction?by?gene-expression?profiling?and?supervised?machine?learning.Nature?Med?2002；8：68-74.

26.Rosenwald?A，Wright?G，Chan?WC，et?al.The?use?of?molecular?profiling?to?predict?survival?after?chemotherapy?for?diffuse?large?B-cell?lymphoma.New?Eng?J?Med?2002；346：1937-1947.

27.Huang?E，Cheng?SH，Dressman?H，et?al.Gene?expression?predictors?of?breast?cancer?outcome.Lancet?2003；361：1590-1596.

28.Ayers?M，Symmans?WF，Stec?J，et?al.Gene?expression?profiles?predictcomplete?pathologic?response?to?neoadjuvant?paclitaxel?and?fluorouracil，doxorubicin，and?cyclophosphamide?chemotherapy?in?breast?cancer.J?Clin?0ncol?2004；22：2284-2293.

29.Mazzanti?C，Zeiger?MA，Costourpus?N，et?al.Using?gene?expression?profiling?to?differentiate?benign?versus?malignant?thyroid?tumors.Cancer?Res2004；64：2898-2903.

30.Roepman?P，Wessels?LFA，Kettelarij?N，et?al.An?expression?profile?for?diagnosis?of?lymph?node?metastasis?from?primary?head?and?neck?squaamous?cell?carcinoma.Nature?Genetics?2005；37：182-186.

31.Cheng?SH，Tsai?SYC，Yen?KL，et?al.Prognostic?significance?of?parapharyngeal?space?venous?plexus?and?marrow?involvement；potential?landmarks?of?dissemination?for?satge?I-III?nasopharyngeal?carcinoma.Int?J?Radiat?Oncol?Biol?Phys?2005；61：456-465.

32.Bolstad?BM，Irizarry?RA，Astrand?M，et?al.A?comparison?of?normalization?methods?for?high?density?oligonucleotide?array?data?based?on?variance?and?bias.Bioinformatics?2003；19：185-193.

33.Lee?AWM，Sze?WM，Au?JSK，et?al.Treatment?results?for?nasopharyngeal?crcinoma?in?the?modern?era：the?hong?Kong?Experience.Int?J?Radiat?Oncol?Biol?Phys?2005；61：1107-1116.

34.Yokota?S，Yamamoto?Y，Shimizu?K，et?al.Increased?expression?of?cytosolic?chaperonin?CCT?in?human?hepatocellular?and?colonic?carcinoma.Cell?Stress?Chaperones?2001；6：345-350.

35.Liu?H，Li?J?&?Wong?L.Use?of?extreme?patient?samples?for?outcome?prediction?from?gene?expression?data.Bioinformatics?2005；21：3377-3384.

36.Simon?R，Radmacher?MD，Dobbin?K，et?al.Pitfalls?in?the?use?of?DNA?microarray?data?for?diagnostic?and?prognostic?classification.JNCI2003；95：14-18.

37.Chang?Y，Lee?TC，Li?JC，Lai?TL，Chua?HH，Chen?CL，Doong?SL，Chou?CK，Sheen?TS，Tsai?CH.Differential?expression?of?osteoblast-specific?factor?2and?polymeric?immunoglobulin?receptor?genes?in?nasopharyngeal?carcinoma.Head?Nec?k.2005Aug?31

38.Fang?WY，Liu?TF，Xie?WB，Yang?XY，Wang?S，Ren?CP，Deng?X，Liu?QZ，Huang?ZX，Li?X，Ding?YQ，Yao?KT.Reexploring?the?possible?roles?of?some?genes?associated?with?nasopharyngeal?carcinoma?using?microarray-based?detection.Acta?Biochim?Biophys?Sin(Shanghai).2005Aug；37(8)：541-6.

39.Lung?HL，Bangarusamy?DK，Xie?D，Cheung?AK，Cheng?Y，Kumaran?MK，Miller?L，Liu?ET，Guan?XY，Sham?JS，Fang?Y，Li?L，Wang?N，Protopopov?AI，Zabarovsky?ER，Tsao?SW，Stanbridge?EJ，Lung?ML.THY1is?a?candidate?tumour?suppressor?gene?with?decreased?expression?in?metastatic?nasopharyngeal?carcinoma.Oncogene.2005Jun?20；[Epub?ahead?ofprint]

40.Zhang?B，Nie?X，Xiao?B，Xiang?J，Shen?S，Gong?J，Zhou?M，Zhu?S，Zhou?J，Qian?J，Lu?H，He?X，Li?X，Hu?G，Li?G.Identification?of?tissue-specific?genes?in?nasopharyngeal?epithelial?tissue?and?differentially?expressed?genes?in?nasopharyngeal?carcinoma?by?suppression?subtractive?hybridization?and?cDNA?microarray.Genes?Chromosomes?Cancer.2003?Sep；38(1)：80-90.

41.Guo?X，Lui?WO，Qian?CN，Chen?JD，Gray?SG，Rhodes?D，Haab?B，Stanbridge?E，Wang?H，Hong?MH，Min?HQ，Larsson?C，Teh?BT.Identifying?cancer-related?genes?in?nasopharyngeal?carcinoma?cell?lines?using?DNA?and?mRNAexpression?profiling?analyses.Int?J?Oncol.2002?Dec；21(6)：1197-204.

42.Hui?AB，Lo?KW，Teo?PM，To?KF，Huang?DP.Genome?wide?detection?of?oncogene?amplificatious?in?nasopharyngeal?carcinoma?by?array?based?comparative?genomic?hybridization.Int?J?Oncol.2002Mar；20(3)：467-73.

43.Soo?R，Putti?T，Tao?Q，Goh?BC，Lee?KH，Kwok-Seng?L，Tan?L，Hsieh?WS.Overexpression?of?cyclooxygenase-2?in?nasopharyngeal?carcinoma?and?association?with?epidermal?growth?factor?receptor?expression.Arch?Otolaryngol?Head?Neck?Surg.2005?Feb；131(2)：147-52

44.Lee?SW，Cho?KJ，Park?JH，Kim?SY，Nam?SY，Lee?BJ，Kim?SB，Choi?SH，Kim JH，Ahn?SD，Shin?SS，Choi?EK，Yu?E.Expressions?of?Ku70?and?DNA-PKcs?as?prognostic?indicators?of?local?control?in?nasopharyngeal?carcinoma.Int?J?Radiat?Oncol?Biol?Phys.2005?Aug?1；62(5)：1451-7.

45.Ma?BB，Poon?TC，To?KF，Zee?B，Mo?FK，Chan?CM，Ho?S，Teo?PM，Johnson?PJ，Chan?AT.Prognostic?significance?of?tumor?angiogenesis，Ki?67，p53oncoprotein，epidermal?growth?factor?receptor?and?HER2?receptor?protein?expression?in?undifferentiated?nasopharyngeal?carcinoma--a?prospective?study.Head?Neck.2003?Oct；25(10)：864-72.

46.Farias?TP，Dias?FL，Lima?RA，Kligerman?J，de?Sa?GM，Barbosa?MM，Goncalves?FB?Jr.Prognostic?factors?and?outcome?for?nasopharyngeal?carcinoma.Arch?Otolaryngol?Head?Neck?Surg.2003?Jul；129(7)：794-9.

47.Hsiao?JR，Jin?YT，Tsai?ST，Shiau?AL，Wu?CL，Su?WC.Constitutive?activation?of?STAT3?and?STAT5?is?present?in?the?majority?of?nasopharyngeal?carcinoma?and?correlates?with?better?prognosis.Br?J?Cancer.2003?Jul?21；89(2)：344-9.

48.Chan?AT，Lo?YM，Zee?B，Chan?LY，Ma?BB，Leung?SF，Mo?F，Lai?M，Ho?S，Huang?DP，Johnson?PJ.Plasma?Epstein-Barr?virus?DNA?and?residual?disease?after?radiotherapy?for?undifferentiated?nasopharyngeal?carcinoma.J?Natl?Cancer?Inst.2002?Nov?6；94(21)：1614-9.

49.Ma?J，Nicholas，Terry?HA，Lin?SX，Patel?N，Mai?HG，Hong?MH，Lu?TX，Cui?NJ，Min?HQ.Prognostic?significance?of?DNA?ploidy?and?proliferative?indices?in?patients?with?nasopharyngeal?carcinoma.Ai?Zheng.2002?Jun；21(6)：644-50.

50.Rubio?L，Burgos?JS，Lopez-Guerrero?JA，Morera?C，Vera-Sempere?FJ.Expression?of?p53?protein?and?tumor?angiogenesis?as?prognostic?factors?in?nasopharyngeal?carcinoma?patients.Pathol?Res?Pfact.2002；198(2)：97-102.

51.Fujii?M，Yamashita?T，Ishigufo?R，Tashiro?M，Kameyama?K.Significance?of?epidermal?growth?factor?receptor?and?tumor?associated?tissue?eosinophilia?in?the?prognosis?of?patients?with?nasopharyngeal?carcinoma.Auris?Nasus?Larynx.2002?Apr；29(2)：175-81.

52.o?YM，Chan?AT，Chan?LY，Leung?SF，Lam?CW，Huang?DP，Johnson?PJ.Molecular?prognostication?of?nasopharyngeal?carcinoma?by?quantitative?analysis?of?circulating?Epstein-Barr?virus?DNA.Cancer?Res.2000?Dec15；60(24)：6878-81.

53.Heng?DM，Wee?J，Fong?KW，Lian?LG，Sethi?VK，Chua?ET，Yang?TL，Khoo?Tan?HS，Lee?KS，Lee?KM，Tan?T，Chua?EJ.Prognostic?factors?in?677?patients?in?Singapore?with?nondisseminated?nasopharyngeal?carcinoma.Cancer.1999Nov?15；86(10)：1912-20.

54.Guo?X，Min?HQ，Zeng?MS，Qian?CN，Huang?XM，Shao?JY，Hou?JH.nm23-H1?expression?in?nasopharyngeal?carcinoma：correlation?with?clinical?outcome.Int?J?Cancer.1998?Dec?18；79(6)：596-600.

Table 1: the formula that is used for the Logic Regression Models of the gene that nine SOM gene clusters select

Affymetrix probe sets reference numbers in the bracket representes to specify the normalized expression intensity of the mRNA transcript of probe sets.From the formula of listing, calculate Lip river base mark (logit score), be used for shifting risk through the prediction of k-near neighbor method.

Table 2: be used for calculating the formula of Lip river base mark (logit score) with the risk of estimation DISTANT METASTASES IN based on 12 expression of gene intensity

The normalized expression intensity that Affymetrix probe sets reference numbers in the bracket representes to specify the mRNA of probe sets to transcribe.Calculate Lip river base mark from the formula of listing, be used to estimate the probability of low DISTANT METASTASES IN risk.Low DISTANT METASTASES IN risk is thought in probability＞0.5.Probability＜0.5 thinks that high distant place shifts risk.

Table 3:138 name patient's Clinical symptoms

*: DISTANT METASTASES IN does not take place all with examining more than 3 years in all clinical low-risk cases.

*: all clinical excessive risk cases first the treatment after 3 years in DISTANT METASTASES IN has all taken place

The parameter of listing does not have statistically-significant difference in training group and testing group patient distribution

Table 5: through 12 genes of logistic regression method prediction DISTANT METASTASES IN

Adding bright gene does not exist in table 4.

Table 6: based on the consistance that predicts the outcome of 52 genes and 12 genes

*: " 0 " is meant low DISTANT METASTASES IN risk; *: " 1 " is meant that high distant place shifts risk; * *: " 9 " are meant inconsistent between two kinds of Forecasting Methodologies.

It is for referencial use that the full text of all patented claims, patent and publication that this paper quotes is incorporated this paper at this.

Previous embodiment can obtain similar achievement through changing used in this embodiment general or special reactant and/or the operating conditions of describing of the present invention.

From the description of preamble, those skilled in the art can be easy to understand essential characteristic of the present invention, under the prerequisite that does not depart from spirit and scope of the invention, can carry out various changes and revise to adapt to its various uses and condition the present invention.

Claims (1)

1. assess the method that the risk of DISTANT METASTASES IN takes place the nasopharyngeal carcinoma patient for one kind; Said method is included in 52 expression of gene spectrums listing from evaluation form 4 in said patient's the sample; The prediction rule that said assessment is to use the k-arest neighbors sorting technique through the group that is applied to Lip river base mark (logit score) to produce is carried out, and the group of said Lip river base mark is to produce to each gene cluster in 9 gene clusters shown in the table 4 through corresponding regression model formula in the use table 1; Use 12 expression of gene spectrums listing in the following prediction rule evaluation form 5,

Wherein said Lip river base mark (logit score) is to use the formula of regression model shown in the table 2 to produce based on said 12 expression of gene spectrum; Contrast derives from the DISTANT METASTASES IN value-at-risk of said two prediction rule; Be when low or high when the DISTANT METASTASES IN risk of confirming through these two kinds of methods thus; Then be recorded as low or excessive risk respectively, when said definite risk was inconsistent, it was uncertain then to be recorded as said risk.

Images